JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 315 Direct Determination of Lead in Sea-waters by Laser-excited Atomic Fluorescence Spectrometry* Venghout Cheam Josef Lechner Ivan Sekerka and Roland Desrosiers Environment Canada National Water Research Institute Research and Applications Branch Burlington Ontario Canada L7R 4A6 This paper describes a laser-excited atomic fluorescence spectrometric method for direct determination of lead in sea-waters down to femtogram levels. No separation/concentration steps nor chemical modifier were used. A programmable in situ known addition technique was developed and used as an integral part of the method. The technique reduces sample preparation steps and compensates for spectral line drift better than a standards calibration technique.Four sea-water Certified Reference Materials from the National Research Council of Canada were analysed for Pb concentrations which were found to be well within certified ranges. Spike recoveries of 100 _+ 10% were achieved using a Certified Reference Material and an unknown sea-water sample. The practical detection limit was 3 fg of Pb absolute (or 1 ng I-' relative) which to the authors' knowledge is the lowest absolute detection limit ever reported for sea-water analysis. Keywords Laser-excited atomic fluorescence spectrometry; sea-water; lead; known addition; ferntogram detection limit Sea-water is a very complex matrix containing trace levels of lead and other elements (ngl-') which act as interferents. These characteristics form a troublesome but challenging prob- lem for analytical chemists as evidenced in the literature some of which are cited here.'-27 The bulk of the past work relies on means to discard the salt matrix and preconcentrate the metals before their determination using various methodologies.The most common methodology used has been electro- thermal atomic absorption spectrometry ( ETAAS).5,6,9,12,14,16 Other common methodologies include isotope dilution mass spectrometry ( ID-MS),9~"~'5*18~20~27 inductively coupled plasma mass spectrometry (ICP-MS) by standards addition or c a l i b r a t i ~ n ' ~ ' ~ ~ * ~ ~ and anodic stripping voltametry (ASV).8i'4 None of these methods is sensitive enough to detect ng 1-' concentrations of lead in sea-water without time- consuming preconcentration.The most common sample pre-treatment has been the use of off-line separation/concentration techniques. Chelating resins such as Chelex- 1003*6*9,14927 and Chelamine25 have often been used. Another common approach has been the chelation-extraction technique using carbamate5 ammonium pyrrolidin-1-yldithioformate-isobutyl methyl ketone (APDC- IBMK)? dithizone-chloroform," APDC-diethylammonium diethyldithiocarbamate (APDC-DDDC),I4 or APDC-silica gel C18.23 Other off-line separation/concentration techniques use immobilized chelates 8-hydroxyquinoline ( or poly-5-vinyl-8-hydroxyquinoline,4 c~precipitation;~.'~ elec- trochemical deposition on mercury e l e c t r ~ d e ; ~ . ~ ~ evaporation to dryness;' and reductive precipitation." The in situlon-line separation/concentration technique is less commonly used volatile hydrides,16 volatile tetraethyllead,2' and 8-hydro~yquinoline.~~~~~ Most of these techniques are tedious and the various sources of contamination must be closely monitored to achieve a reasonably low blank concentration compared with sample concentration." Table 1 shows that some reported Pb blanks are as high or even higher than sea- water concentrations which can be as low as l ng l-'.(ref. 22). The best blank value reported appears to be 3.7 pg of Pb.24 Although validated analytical methods based on laser- excited atomic fluorescence spectrometry (LEAFS) are few there have been numerous applications of LEAFS to the analyses of real substrates. These include for example river ~ a t e r s ~ ~ - ~ O snow and i ~ e ~ ~ - ~ ~ tap water,35 blood,36 Great 17,19,21,23-25 * Presented at the XXVIII Colloquium Spectroscopicurn Inter- nationale (CSI) York UK June 29-July 4 1993.Lakes w a t e r ~ ~ ~ ~ ~ ore-process solutions,39 biological and vegetation sample^,^@'^ air,43 soil and sediment sample^,^'.^^ and pure metals.40.42,45 Ho wever there has been no method developed for analysis of sea-waters. This paper describes the first LEAFS method for direct determination of femtogram levels of lead in sea-water. No separation/concentration steps nor chemical modifier were necessary. As the sea-water matrix has a wide range of salinity and there is no representative background matrix the stan- dards calibration technique cannot be used with reliable accu- racy. A programmable in situ known addition (standard addition) technique has been developed and used as an integral part of the method. For 3 p1 of sea-water the practical detection limit was 3 fg of Pb absolute (or 1 ng 1-1 relative).Experimental LEAFS The LEAFS instrumentation has been described e l s e ~ h e r e . ~ ~ ? ~ ~ Some key features are given here. The 511 nm line of a copper vapour laser (Metalaser Technologies MLT20) was used to optically pump a Rhodamine 6G dye laser (Laser Photonics). The dye laser output (566 nm) was then frequency-doubled by a second harmonic generator (Autotracker 11 Inrad I) to give the 283 nm ultraviolet light. This light directed through a pierced mirror into a graphite furnace (Perkin-Elmer HGA 2100) was used to excite Pb atoms generated in the furnace.The fluorescent light (406nm) emitted by the excited atoms was collected and measured via a monochromator-photomulti- plier-boxcar system. A 6 kHz repetition rate was used and the peak irradiance of the 283 nm beam in the furnace was about 2 kW cmP2. Sample Handling Sample handling was carried out in a class 100 clean room and in a class 100 laminar flowhood (Microzone Corporation) except for sample injection (25 p1 total liquid volume) which was carried out in a regular laboratory atmosphere. Details of labware cleaning procedure and clean room development have been described earlier.46 Ultrapure chemicals were used. Milli-Q water (Millipore) acidified to 0.2% with ultrapure nitric acid (Seastar) simply referred to as MQW was used as the standards matrix.316 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 Table 1 Some reported Pb blank concentrations for various preconcentration techniques Technique Chelation-Extraction Carbamate extraction Di thizone-chloroform Chelamine Chelex- 100 resin Reductive precipitation Tetrahydroborate Immobilized chelate 8-H ydroxyquinoline 8-H ydrox yquinoline Hg film electrode Resin Electrochemical deposition In situion-line concentration I-8-HOQ Methodology ETAAS IDMS ETAAS ICP-MS ETAAS ETAAS ID-ICP-MS DP- ASVt FI-ETAAS Pb blank concentration Ref. 5 24-36 ng 1-' total blank 0.1 ng per extraction 11 1.2 ng absolute 25 0.08 ng column blank 27 1.7 ng absolute 19 <0.8 ng absolute 12 2.99 ng 1-' final column blank* 18 z55 ng I-' 14 3.7 pg absolute 24 *Various blanks ranging from 0.2 to 140 ng 1-'.TDP-ASV = differential pulse anodic stripping voltammetry. Table 2 Results of replicate weighings of 3 and 5 pl of solutions Parameter n 5 p l O f MQW- MeanICLg s/Pg RSD (%) 3 pl of sea-water- n MeaniCLg SiPg RSD (Yo) Run 1 9 4.94 0.2 4.07 9 3.12 0.13 4.21 Run 2 9 5.05 0.11 2.2 9 3.09 0.06 1.82 Run 3 9 5.05 0.16 3.2 9 3.05 0.09 :!.89 Run 4 9 5.02 0.18 3.64 9 3.11 0.09 2.95 Run 5 9 4.99 0.13 2.58 9 3.11 0.09 2.9 Overall 45 5.01 0.16 3.16 45 3.1 0.09 3.05 Programmable Micropipette An electronic micropipette (Rainin Instrument Company motorized microliter pipette Model EP-100) was used to sample microlitre volumes of different liquids or air (as separate segments) into the same tip and to inject the whole into the graphite tube.Tapered pipette tips were found to be necessary for reliable sampling and injection. The tips were acid-soaked rinsed and tested for very low blanks before use. The segmented sampling of 17 pl of MQW (carrier) followed by 3 pl of sea- water 2 p1 of air (spacer) and 5 pl of standard (or MQW) was used. The standard concentration used varied from 5 to 50 ng 1-l. Peak heights were used to calculate Pb concentration in sea-water by the known addition procedure. Results and Discussion The LEAFS optimization has been described e l ~ e w h e r e . ~ ~ . ~ ~ Choice of Purge Gas A brief review of the literature indicates several advantages of adding H to the Ar purge gas in electrothermal techniques. Amos et al.47 observed a 20-fold improvement in sensitivity for the atomic absorption of aluminium when H2 was intro- duced into the Ar carrier gas.They also showed that both spectral and chemical interferences (caused by excess of such interferents as H,PO NaCl KCl MgCl and CaC1,) were considerably reduced when the Ar-H mixture was used for the determination of Pb by atomic absorption spectrometry. Schmid and Krivan4* proved that in 1% NaCl matrix the stability of Pb in the graphite tube improved with the addition of H2 to Ar. For tungsten or other metal atomizers the usual purge gas is Ar mixed with H2 to provide a reducing environ- ment and to prolong the atomizer lifetime.49~50 Goforth and Winefordner4 compared Ar to an Ar-H mixture in their LEAFS study and found that the Ar-H mixture gave better detection limits than Ar for several elements although no comparison was made for Pb.In this study it was found that the Pb fluorescence signals were improved by about 35% if an Ar-H mixture (92-8%) instead of pure Ar was used as the furnace purge gas for our LEAFS system. Furthermore this gas mixture gave better stability in fluorescence signals and a longer tube lifetime ( z 200 firings). Obviously the mixture provides a superior environment than pure Ar for this work probably because of the reducing environment generated by the burning of H2 during the atomization step.47 Furnace Conditions Uncoated graphite tubes were used as they provide an adequate rise time during the atomization period; the pyrolytic graphite coated graphite tubes give too slow a rise time for the Perkin-Elmer HGA 2100.The atomization temperature also depends on tube thickness for example 2100°C for 1 mm and 2300 "C for 1.3 mm thickness. The optimal drying-pyrol- ysis-atomization temperatures and times were as follows 120 "C for 40 s ramped 350 "C for 40 s unramped and 2300 "C for 3 s unramped. The purge gas flow was stopped during atomization. Pipette Accuracy Normal injections of 20 p1 of sample resulted in too much salt matrix in the furnace tube and subsequent signal suppression so a smaller volume was used. Even though the manufacturer's recommended volume range for the micropipette is 10-100 pl a weighing experiment showed that 3 p1 of sea-water can be reliably picked up and dispensed taking the density of sea- water to be 1.025 g ml-'. The same performance was achieved with 5 pl of MQW as shown in Table 2 hence the 3 and 5 pl volumes were taken as true.The weighing experiment wasJOURNAL O F ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 317 2.0 Salinity 0 1.2% m 3.5% 1.5 3 c" 1.0 cn 0 .- 0.5 o . . . . . . . . . . . . n . . 0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 S a rn ple vo I u rn e/pl Fig. 1 Signal u e r m sample volume for low and high salinity samples repeated for 2 pl of sea-water but the reproducibility was inferior to that obtained with 3 pl. Effect of Water Salinity Two water samples of low and high salinity (12 and 35%0) were used to study the effect of salinity on the fluorescence signal. To provide a constant volume of injected solution a volume of 25 pl (the sum of sea-water and MQW volumes) was chosen for a series of solutions containing an increasing amount of sea-water.Fig. 1 depicts the fluorescence signal versus the volume of sea-water. It shows that for the highly saline water there is a small range of maximum response peaking around 3 pl of sea-water. For the low salinity water the range of maximum response is relatively wide commencing at about 3 pl of sea-water. Since the two salinity values cover the range usually encountered 3 p1 of sea-water was taken as the optimal volume. Sampling Sequence and In Situ Known Addition Analysis The manual injection of 3 pl of sea-water alone into the graphite furnace resulted in very irreproducible responses possibly caused by the irreproducible delivery of such a small amount through the furnace tube sample hole and the irreprod- ucible distribution of solution on the tube itself thus resulting in a variable position of the atomic cloud with respect to the small laser beam.To help attain reproducible sample injection into the graphite tube the inclusion of a carrier (MQW) was necessary. Also as there is no representative background matrix normal standards calibration could not be used and a known addition technique was relied upon which has the advantage of minimizing the effect of instrumental or baseline drifts that the standards calibration technique often encounters. For example the dye laser retuning was clearly less frequent with the known addition technique. To adequately encompass the sample carrier and standard into one containment without premixing the individual solutions a sampling sequence pro- cedure was devised using a programmable micropipette.The technique comprises segmented sample pick-ups and the dis- pensing of the whole into the graphite tube. A number of different sequences of sample pick-ups were tried some of which are (i) pick-up of 5 pl of standard (or MQW) followed by 2 pl of air 3 p1 of sample 2 pl of air and 17 pl of MQW carrier; (ii) sequence (i) followed by an additional step that of picking up 71 pl of air in order to mix the different liquid fractions; (iii) sequence ( i ) in reverse order; (iv) pick-up of 5 pl of standard (or MQW) followed by 2 pl of air 2 yl of sample and 18 pl of MQW carrier; (u) sequence (iv) followed by 73 pl of air to mix the different fractions; and (vi) sequence (iv) in reverse order.The optimal sequence was the pick-up of 17 pl of MQW followed by 3 pl of sea-water 2 pl of air spacer and 5 pl of standard (or MQW); this is referred to as sequence 17-3-2-5 1 2 3 4 5 t t . . t i Or standard) - I Fig. 2 Pick-up and dispense sequence 17-3-2-5 1 pick-up carrier (17 pl MQW); 2 pick-up sea-water (3 p1); 3 pick-up air spacer (2 pl); 4 pick-up standard or MQW (5 pl); and 5 dispense total content of tip. Injected volume (sample + MQW) = 25 pl which is schematically depicted in Fig. 2 showing details of the various steps. The carrier MQW as well as effectively delivering the different segments into the furnace also serves to rinse the pipette tip for next use. A determination consists of analysing the sequence containing 5 pl of MQW followed by the sequence containing 5 pl of standard.A computer program was written and used to calculate sea-water sample concentration and is available on request. The complex sea-water matrix may generate background signals which interfere with the analyte fluorescence signal. Three types of background exist concomitant scatter molecu- lar fluorescence and non-analyte atomic fl~orescence.~~ To determine these backgrounds signals were obtained for sea- water samples at k0.05 nm away from the analytical line and were found to be smaller than the furnace blank observed at the analytical line indicating no background interference. This supports previous that Pb background signals for electrothermal LEAFS are negligible compared with the fluor- escence signal. In situ known addition analyses of MQW yielded 0.42f0.17 ng 1-1 of Pb based on 18 determinations made on two different days.This result supports a previously reported value of <0.9 ng 1-1 (ref. 27). Performance Indicators Fig. 3 shows the typical fluorescence responses for analysis of real samples unspiked and spiked at various concentration levels. The two samples PC-1 and PC-2 with salinity of 31 and 22%0 were collected from the near shore of the Pacific Ocean at Burrard Inlet (North Vancouver) and at Atchinson Point (West Vancouver). The duplicate analyses of the 17 spiked and unspiked samples produced an RSD range of 0.3-8.4% with an average RSD of 2.5%. Calibration curves for this technique are linear over at least three orders of magnitude which is more than adequate for this work.Although the complete linear dynamic range was not obtained as it was unnecessary there is no reason why it should not extend over four or five orders of magnitude. Based on the definition of the International Union of Pure and Applied Chemistry (IUPAC) the detection limit was equated to three times the standard deviation sb of eighteen determinations of the blank (3 yl of MQW) and was found to be 0.5 ng l-l corresponding to 1.5 fg absolute. It is worth noting that replicate analyses of 25 pl of MQW using standards calibration gave a 3s value of 0.6 ng l-' which is practically the same as the 3sb. In principle as the sample matrix effect is cancelled by the known addition technique the effective blank should be the furnace blank; 25 furnace blank responses were taken over a period of several weeks and calculations made318 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 r I - x 03 + Sample Fig. 3 Typical fluorescence signals for spiked and unspiked sea-water samples using the 17-3-2-5 sequence; FB =furnace blank and PC-2 (0) = unspiked sea-water sample from Pacific Ocean (values in parentheses represent amount of Pb in ng 1-'). Boxcar sensitivity was set at 2 V to the left of the vertical line and 1 V to the right of the line. based on concurrent analyses of 16 different sea-water samples giving a 3sb value equal to 0.8 ng1-'. Consequently the detection limit was taken as 1 ng I-' (3 fg absolute). This limit was further confirmed by replicate analyses of a sea-water sample containing 6ngld1 resulting in a 2s value equal to 0.9 ng 1-' which would be the usual definition of a working detection limit.52 The femtogram detection limit for sea-water analysis is to our knowledge the lowest so far achieved (Table 3).Several unspiked and spiked natural samples of various origins and salinities were used for a test of accuracy. Four Certified Reference Materials from the National Research Council of Canada (NASS-4 NASS-3 CASS-3 and SLEW-1) Table 3 Comparison of some recent detection limits were analysed. Table 4 shows that the LEAFS results compare well with the certified values. Also NASS-4 and PC-2 were spiked at 3 different levels and analysed showing adequate recoveries (Table 5). Analysis of a Pacific Ocean Profile Sea-water samples of a vertical profile were collected from a location in the Pacific Ocean between Japan and Hawaii aboard a Russian research vessel.The sampling location was 27" 47' N 174" 59' E and is referred to as AV-HS 10 (Aleksandr Vinogradov Hydro Station No. 10). Sampling details have been provided by Orians and Yang.53 The samples were Methodology DP-ASV* ETAAS-Extraction ETAAS-Chelex ICP-MS$ ICP-MS Flow injection-ETAAS LEAFS Preconcentration procedure Rotating glassy carbon Hg film electrode Chelex- 100 Chelex-100 None APDC-DDDC I-8-HOQ I-8-HOQ Original sample volume/ml 50 250-300 2000 500 1000 5 0.003 Relative detection limit/ng I-' 1 Absolute detection limit/pg 50 5 1.2 0.4 0 4 1.14 1 20T 247 20 5 q 5.7 0.003 Ref. 14 14 14 15 27 24 This work *DP-ASV = differential pulse anodic stripping voltammetry. tAssuming the usual injection volume of 20 pl was used each time.$Standard additions ICP-MS; 3 aliquots of 500 ml of sea-water were used. §Instrumental detection limit =0.03 pg 1-' (ppb) (ref. 53). TAssuming 1 ml of preconcentrated sea-water is aspirated into the ICP mass spectrometer. Table 4 LEAFS results uersus certified values for Pb in certified reference materials Certified Found/ng 1-l No. of Material Salinity (%o) Origin values/ng 1- ' f SD determinations NASS-4 31.3 North Atlantic 13-1-5 11.4 1.5 30 open ocean open ocean nearshore River estuary NASS-3 35.1 North Atlantic 39+6 40.5 f 3.0 12 CASS-2 29.2 North Atlantic 19k6 18.4+ 1.9 12 SLEW-1 11.6 St. Lawrence 28k7 30.2 f 3.0 10JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 319 Table 5 Recoveries (%) of spiked and unspiked natural samples Concentration Recovery No. of Sample Salinity (%o) Origin range/ng 1-' (%I determinations Spiked 31.3 North Atlantic 11-41 102+ 10 15 NASS-4* open ocean Unspiked 21.9 North Pacific 16.4 100 15 PC-2 near shore PC-2* nearshore Spiked 21.9 North Pacific 16-46 102 f 6 12 *Spiked at three different levels of 10 20 and 30 ng 1-l; 4 or 5 replicate analyses at each level. Table6 Pb concentrations of a vertical profile of Pacific Ocean [location AV-HS 10 (27" 47' N 174" 59' E)] Depth/m Concentration/ng 1- ' ICP-MS LEAFS* 25 13.2 14.8 f 1.4 75 13.4 16.8 f 1.1 250 15.1 17.5 f 0.1 500 12.5 16.4f 1.6 lo00 9.1 7.6 f 0.6 2 500 6.6 5.2 f 0.2 ~ ~~ ~ ~~~~~ *Value t- standard deviation of two determinations. analysed by the LEAFS method and the results compare well with those by ICP-MSS3 (Table 6).Recently published data for the Pacific and Atlantic seem to indicate similar Pb levels and profiles between the Oceans and the Great Lakes waters (particularly Lake Superior).38 Conclusion A LEAFS method has been developed for the direct determi- nation of femtogram levels of lead in sea-waters. The in situ known addition technique reduces sample preparation steps and hence possible contamination sources. It also compen- sates better for analytical line and instrumental drifts than the standards calibration technique and is a potential standard procedure using LEAFS for trace metal determination in waters of diverse matrices. The proposed LEAFS method also lends itself to easy automation and robotization. We acknowledge Professor Kristin Orians of the University of British Columbia for kindly providing profiles samples data and information on the ICP-MS method. We also acknowledge Greg Lawson for meticulous sample preparations.Thomas- Louis Tremblay and John Kraft for collecting some sea- water samples. 1 2 3 4 5 6 7 8 9 10 References Young E. G. Smith D. G. and Langille W. M. J. Fish. Res. Board Can. 1959 16 7. Weiss H. V. and Lai M. G. Talanta 1961 8 72. Riley J. P. and Taylor D. Anal. Chim. Acta 1968 40 479. Buono J. A. Karin R. W. and Fashing J. L. Anal. Chim. Acta 1975 80 327. Danielsson L.-G. Magnusson B. and Westerlund S. Anal. Chim. Acta 1978 98 47. Kingston H. M. Barnes I. L. Brady T. J. Rains T. C. and Champ M. A. Anal. Chem.1978 50 2064. Guevremont R. Sturgeon R. and Berman S. Anal. Chim. Acta 1980,115 163. Halliday M. C. Houghton C. and Ottaway J. M. Anal. Chim. Acta 1980 119 67. Sturgeon R. E. Berman S. S. Desaulniers J. A. H. Mykytiuk A. P. McLaren J. W. and Russel D. S. Anal. Chem. 1980 52 1585. Guevremont R. Anal. Chem. 1981 53 911. 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 Shaule B. K. and Patterson C. C. Earth Planet. Sci. Lett. 1981 54 97. Sturgeon R. E. Berman S. S. Willie S. N. and Desaulniers J. A. H. Anal. Chem. 1981 53 2337. Grobenski Z. Lehmann R. Radziuk B. and Voellkopf U. At. Spectrosc. 1984 5 87. Bruland K. W. Coale K. H. Mart L. Mar. Chem. 1985 17 285. McLaren J. W. Mykytiuk A. P. Willie S. N. and Berman S .S. Anal. Chem. 1985 57 2907. Sturgeon R. E. Willie S. N. and Berman S . S. Fresenius' Z . Anal. Chem. 1986 323 788. Akatsuka K. and Atsuya I. Fresenius' 2. Anal. Chem. 1987 329 435. Beauchemin D. McLaren J. W. Mykytiuk A. P. and Berman S . S. J. Anal. At. Spectrom. 1988 3 305. Nakashima S. Sturgeon R. E. Willie S. N. and Berman S . S. Anal. Chim. Acta 1988 207 291. Beauchemin D. and Berman S. S. Anal. Chem. 1989,61 1857. Sturgeon R. E. Willie S. N. and Berman S. S. Anal. Chem. 1989 61 1867. Burton J. D. and Statham P. J. in Heavy Metals in the Marine Enuironment eds. Furness R. W. and Rainbow P. S. CRC Press Boca Raton 1990 ch. 2 pp. 6-25. Liu Z.-s. and Huang S.-d. Anal. Chim. Acta 1992 267 31. Azeredo L. C. Sturgeon R. E. and Curtius A. J. Spectrochim. Acta Part B 1993 48 91.Blain S. Appriou P. and Handel H. Anal. Chim. Acta 1993 272 91. Lan C.-r. Analyst 1993 118 189. Yang L. and Orians K. J. University of British Columbia Vancouver Canada personal communication 1993. Epstein M. S. Bradshaw J. Bayer S. Bower J. Voigtman E. and Winefordner J. D. Appl. Spectrosc. 1980 34 372. Epstein M. S. Bayer S. Bradshaw J. Voigtman E. and Winefordner J. D. Spectrochim. Acta Part B 1980 35 233. Remy B. Verhaeghe I. and Mauchien P. Appl. Spectrosc. 1990 44 1633. Bolshov M. A. Kolosnikov V. G. Rudnev S . N. Boutron C. F. Gorlach U. and Patterson C. C. J. Anal. At. Spectrom. 1992 7 99. Bolshov M. A. Boutron C. F. Ducroz F. M. Gorlach U. Kompanetz 0. N. Rudniev S. N. and Hutch B. Anal. Chim. Acta 1991 251 169. Boutron C. F. Bolshov M.A. Kolosnikov V. G. Patterson C. C. and Barkov N. I. Atmos. Environ. 1990,24A 1797. Bolshov M. A. Boutron C. F. and Zybin A. V. Anal. Chem. 1989,61 1758. Falk H. Paetzold H.-J. Schmidt K. P. and Tilch J. Spectrochim. Acta Part B 1988 43 1101. Omenetto N. Human H. G. C. Cavalli P. and Rossi G. Analyst 1984 109 1067. Cheam V. Lechner J. Sekerka I. Desrosiers R. Nriagu J. and Lawson G. Anal. Chim. Acta 1992 269 129. Cheam V. Lechner J. Desrosiers R. Sekerka I. Nriagu J. and Lawson G. Int. J. Environ. Anal. Chem. 1993 53 13. Bolshov M. A. Zybin A. V. and Smirenkina I. I. Spectrochim. Acta Part B 1981 36 1143. Hendrick M. S. Goliber P. A. and Michel R. G. J . Anal. At. Spectrom. 1986 1 45. Dougherty J. P. Costello J. A. and Michel R. G. Anal. Chem. 1988 60 336. Goforth D.and Winefordner J. D. Anal. Chem. 1986 58 2598. Liang Z. Wei G.-T. Irwin R. L. Walton A. P. and Michel R. G. Anal. Chem. 1990,62 1452.320 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 44 Irwin R. L. Wei G.-T. Butcher D. J. Liang Z. Su E. G. Takahashi J. Walton A. P. and Michel R. G. Spectrochim. Acta Part B 1992 47 1497. 45 Dashin S. A. Mayorov I. A. and Bolshov M. A. Spectrochim. Acta Part B 1993 48 531. 46 Nriagu J. O. Lawson G. Wong H. K. T. and Azcue J. M. J. Great Lakes Res. 1993 19 175. 47 Amos M. D. Bennett P. A. Brodie K. G. Lung P. W. Y. and Matousek J. P. Anal. Chem. 1971 43 211. 48 Schmid W. and Krivan V. Anal. Chem. 1985 57 30. 49 Suzuki M. and Ohta K. Prog. Anal. At. Spectrosc. 1983 6 49. 50 Kinrade J. D. Kostiak W. and McCarthy G. L. in Applications Manual Scintrex AAZ-2 Zeeman Modulated Absorption Spectrophotometer Scintrex Toronto 1986 Technical note TN-3. 51 Butcher D. J. Irwin R. L. Takahashi J. Su G Wei G-T. and Michel R. G. Appl. Spectrosc. 1990 44 1521. 52 Cheam V. Analyst 1992 117 1137. 53 Orians K. J. and Yang L. University of British Columbia Vancouver Canada personal communication 1993. Paper 31038931 Received July 6 1993 Accepted September 17 1993